In partnership with

We track Product so you don't have to. Top Podcasts summarised, the latest AI tools, plus research and news in a 5 min digest.

Hey Product Fans!

Welcome to this week’s 🌮 Product Tapas!

If you’ve been forwarded this or just stumbled upon it, you’re in for a treat. For the best reading experience, check out the web version and sign up for future editions here.

What’s cooking this week? 🥘 

📰 Not Boring – Another week, another stack of outrageous new LLM models. OpenAI's dropping new research tools, DeepSeek's making waves, and Gemini 2.0 is claiming the crown. Plus, we've got deep fakes getting scary-good, Apple hitting major milestones (while quietly shelving AR glasses), and Jack Dorsey's entering the AI Agent arena with "Goose".

⌚ Productivity Tapas – Some genuine 100% must-try recommendations this week. Meet Chorus, your new desktop AI command centre, plus Tana's finally out of beta! We're also checking out FounderPal for AI marketing and Granola for next-level meeting notes.

🍔 Blog Bites – Navigate "The Great Product Reset of 2025", master your GTM strategy from scratch, and get back to UX basics with the proximity principle. Three meaty reads to sink your teeth into!

🎙️ Pod Shots – Colin Matthews breaks down his #1 hack for building apps with AI, perfect timing after last week's Bolt.new deep dive!

Plenty to get stuck into - off we go! 🚀

There’s a reason 400,000 professionals read this daily.

Join The AI Report, trusted by 400,000+ professionals at Google, Microsoft, and OpenAI. Get daily insights, tools, and strategies to master practical AI skills that drive results.

Sign up now for free and work smarter, not harder.

📰 Not boring

• OpenAI unveils a new ChatGPT agent for ‘deep research’

  • Hands on with OpenAI’s Deep Research (spoiler alert, it’s awesome)

• OpenAI also hits back at DeepSeek with o3-mini reasoning model

• DeepSeek has been inevitable and here's why (history tells us)

• Gemini -2.0 Pro takes #1 spot across all categories. Literally every week another model comes out that beats the competition 🤯 

• Ingesting millions of PDFs and why Gemini 2.0 changes everything. TL;DR Gemini Flash 2.0 achieves near-perfect OCR accuracy while being incredibly cheap

• Wrappers may have the last laugh: AI-powered code editor Cursor is now reportedly the fastest app in history to go from $1M to $100M in revenue

• Deepfake videos are getting shockingly good - Bytedance releases OmniHuman-1. Here’s a fictional Taylor Swift performance. Here’s a TED Talk that never took place.

• Whilst we’re on models - WTF is going on with model naming? Marketeers, please send help!

  • If OpenAI ran Starbucks:
    • Coffee
    • Coffee-mini
    • Coffee3
    • Coffee3-pro-shot
    • Coffee-pro-mode (extra foam, but no one knows why)
    • Latte
    • Latte-mini-foam
    • Latte3-espresso-boost
    • Cappuccino (beta)
    • Mocha3-ultra (only available to select users)
    • Frappuccino (discontinued but still listed)
    • Cake-pop (alignment-tuned)
    • PumpkinSpice-4-turbo (seasonal, but always leaks early)
    • Mobile-order(buggy)

  • Hat Tip to Dave Killeen for the above

• Apple tops 1 billion subscriptions, nearly $100B in services revenue in 2024

  • But Apple also reportedly gives up on its AR video glasses project

• Adobe’s new AI tool can help decode lengthy contracts

• Snap unveils AI text-to-image model that runs locally on mobile devices

• Figure (humanoid robotics business) drops OpenAI in favour of in-house models

• Amazon's AI revamp of Alexa assistant nears unveiling (FINALLY)

• Jack Dorsey (Twitter co-founder) returns with open-source AI agent “Goose”

• If the internet becomes AI, does advertising become a matter of persuading agents?

• Threads now has 320m MAUs and 100m DAUs. (Twitter reported 240m ‘monetisable DAUs’ in the summer of 2022, just before the acquisition)

• The X Everything App will launch this year, aiming to compete with Apple Pay

• Stablecoins are finding product-market fit in emerging markets

• Mark Zuckerberg teases a 2025 return to ‘OG Facebook’

⌚️ Productivity Tapas: Time-Saving Tools & GPTs

• Chorus: Chat with a bunch of AIs on your desktop. Run all the frontier models, including GPT-4o, Claude Sonnet 3.5, Gemini Flash, Deepseek, OpenRouter, alongside local open-source models running on your own machine. This is wild. Try it now.

• Tana: FINALLY out of beta - AI native workspace to stay on top of everything. So cards on the table; I’ve dabbled in and out of this for a month or two now and struggled. I need to give it some more time though as I genuinely feel it’s HUGE productivity boost if you get the upfront set up right (which I appreciate is a bit of an oxymoron).

• FounderPal: AI marketing platform to grow your business. Understand your marketing strategy, delegate tasks to your AI assistant, fix your conversion rate, and get more high-quality traffic.

• Granola: In a sea of meeting note transcribers, this has risen to be my absolute favourite. Take notes and Granola will listen along and augment them once finished. Works for IRL meetings and zooms (etc.) alike. It just works.

: Remember, as a Product Tapas Pro subscriber you can access the full time saving tools database for fast approaching 400 time-saving tools relevant for product managers and founders 🔥.

Click this link to access.

🍔 Blog Bites - Essential Reads for Product Teams

Learn: The Great Product Reset of 2025

Unless you’re living under a rock, you’ll know that the product management field is experiencing a seismic shift due to AI advancement, creating both challenges (job uncertainty, investment hesitation) and opportunities (new toolkits, evolved roles). As ever, success requires adapting to uncertainty while maintaining focus on customer value.

This article tries to cover how you can best up-skill and navigate the new AI-driven landscape.

GTM: How to build your GTM strategy from scratch

Maja Voje is probably my favourite GTM ‘influencer’ of the moment. In this recent Kyle Poyer growth unhinged article, she covers her framework-first approach to market success.

Successful GTM strategy involves three key frameworks: developing an Early Customer Profile before ICP, creating differentiated positioning through UVP/USP, and building a systematic GTM masterplan. Success requires methodical progression through Problem-Solution Fit, Product-Market Fit, and GTM Fit stages.

“The holy grail is to find at least one repeatable and scalable GTM motion that will drive your growth." It's not about one-off launches but finding sustainable, repeatable paths.

UX Basics: Proximity Principle

The proximity principle, part of Gestalt psychology, states that objects close to each other are perceived as related. In UI design, this principle helps create intuitive interfaces by grouping related elements together and separating unrelated ones. 🤓 

Understanding proximity in UI design is crucial because it directly affects how users process information and interact with interfaces. Poor implementation can lead to confusion and errors, while proper use can enhance usability and user satisfaction.

Yolquin Alimo explains more here.

🎙️ Pod Shots - Bitesized Podcast Summaries

Remember, Product Tapas Pro subscribers get access to an ever growing database of all the top Podcast summaries we’ve ever created. 

Check it out here.

🤖 The #1 Hack to Build Apps With AI Without Getting Stuck | Colin Matthews

In last week’s productivity tapas deeper dive, we explored Bolt, a tool that simplifies workflows and boosts efficiency. Check it out here if you missed it.

This week, we’re again diving into the recent Peter Yang podcast with Colin Matthews on how to build apps with AI. Colin, a seasoned instructor who has trained over 1,000 product managers to become more technical, shares his insights on leveraging AI tools to build apps quickly and effectively.

We cover Colin’s approach, including his reflection technique, the best tools for prototyping, and strategies to avoid common pitfalls when working with AI. Whether you’re a PM, designer, or just curious about AI-driven app development, it’s a great episode packed with useful tips and tricks.

Oh and check out the link at the end. I’ve just found an awesome looking tool to partner with Bolt to further supercharge your AI driven app building 🔥.

 🎥Watch the full episode here

📆 Published: January 26th, 2025

🕒 Estimated Reading Time: 3 mins. Time saved: 37 mins🔥

🪞 The Reflection Technique: Plan Before You Code

Colin’s #1 hack for building apps with AI is a technique he calls reflection. Instead of jumping straight into coding, he recommends:

1. Providing all the necessary information upfront: Clearly describe what you want to build, including features, data models, and user interfaces.

2. Asking for a plan first: Before generating any code, ask the AI to create a detailed implementation plan.

This approach has two key benefits:

• Preview and modify: You can review the plan and make adjustments before any code is written.

• Learn as you go: By understanding the plan, you gain insight into how the system works, even if you’re not technical.

Colin emphasises that even if you don’t read the plan in detail, the AI will still produce better results by following this structured process.

Key Takeaway for PMs and Founders

Before diving into development, take a moment to reflect and plan. This ensures you’re building the right thing and reduces the risk of costly mistakes later. AI can help you think through the process, even if you’re not technical.

🛠️ Choosing the Right AI Prototyping Tool

With so many AI prototyping tools available, it can be overwhelming to choose the right one. Colin categorises these tools into three main groups:

1. Chatbots (e.g., ChatGPT, Claude)

   • Best for writing simple client-side code.

   • Limited when it comes to server-side functionality or deployment.

2. Cloud Development Environments (e.g., Replit, Bolt, V0)

   • Ideal for building full-stack applications, including client-side, server-side, and database integration.

   • These tools also handle deployment, making them a great choice for most users.

3. Local Development Assistants (e.g., Cursor, WindServe)

   • Useful for professional developers who want to integrate AI into their existing workflows.

   • Requires more technical knowledge to manage code locally and deploy it manually.

For beginners, Colin recommends starting with cloud development environments like Replit or Bolt. These tools strike a balance between ease of use and functionality, allowing you to build and deploy apps without extensive technical expertise.

Key Takeaway for PMs and Founders

Choose tools that match your technical skills and project needs. For most PMs and founders, cloud development environments like Replit or Bolt are the sweet spot for building and deploying apps quickly.

🔧 Tool Highlights: Replit, Bolt, and More

Here’s a quick comparison of some popular tools Colin discussed:

• Replit:

  • Best for data-driven applications.

  • Supports Python libraries like Streamlit for dashboards and internal tools.

  • Includes built-in database hosting and deployment tools.

• Bolt:

  • Highly flexible with strong defaults that can be easily customised.

  • Great for prototyping and iterating on designs.

• Lovable:

  • Lacks a code editor, which can be limiting.

  • Offers integrations with tools like Supabase and GitHub for deployment.

If you’re building both client-side and server-side functionality, tools like Replit and Bolt are excellent starting points.

Key Takeaway for PMs and Founders

Understand the strengths and limitations of each tool. Replit is great for data-heavy apps, while Bolt offers flexibility for prototyping. Pick the tool that aligns with your app’s goals.

📋 Building a Habit-Tracking App: A Step-by-Step Example

To demonstrate his approach, Colin walked through building a simple habit-tracking app using AI. Here’s how he did it:

1. Define the Features:

   • Users can add habits, mark them as completed, and track streaks over time.

   • The app includes two screens: a main screen for habit tracking and a detail view for individual habits.

2. Write a Detailed Prompt:

   • Colin described the user interface (e.g., date navigation, habit completion tracking) and the data model (e.g., habits associated with dates).

   • He also specified future requirements, like adding a calendar view, to ensure the data model was compatible.

   • Specific prompt:

     "I want to build a habit-tracking app. It should allow me to:

     Enter new habits. Mark if I completed the habit. Show me a streak of how many days in a row I’ve completed the habit, from the current date to the past.

     The user interface should include:

     The date at the top, with left and right navigation to change the date. Each habit’s completion status should be associated with a specific date. Habits can be marked as complete or incomplete for each date.

     Future Requirements:

     In the future, I want to add a calendar view. Ensure the data model is compatible with this feature.

     Instructions:

     Start by creating a plan for the implementation, beginning with the data model. Do not write any code yet."

   • Why This Prompt Works

     • Clarity: It clearly defines the app’s purpose and features.

     • Specificity: It describes both the user interface and the data model requirements.

     • Future-Proofing: It anticipates future features (e.g., a calendar view) and ensures compatibility.

3. Ask for a Plan:

   • Before generating any code, Colin asked the AI to create a plan, starting with the data model.

   • This step helped identify potential issues early, such as missing data elements.

4. Iterate and Debug:

   • Colin used an iterative approach, implementing one feature at a time.

   • When errors occurred, he asked the AI to explain the issue before attempting a fix.

5. Deploy the App:

   • The app was deployed using Netlify, a hosting platform integrated with tools like Bolt.

Key Takeaway for PMs and Founders

Start small and iterate. By breaking the app into manageable steps, you can build faster, debug more effectively, and adapt to changes along the way.

🧠 Best Practices for AI-Driven App Development

Colin shared several strategies to avoid common pitfalls when building apps with AI:

1. Nail the Data Model Early:

   • Most issues arise from poorly defined data models. Spend time upfront to ensure your data structure supports future features.

2. Take an Iterative Approach:

   • Instead of asking the AI to build the entire app at once, break it into smaller tasks.

   • This makes it easier to debug and refine each component.

3. Be Specific:

   • The more detailed your prompts, the better the results. Include file names, UI elements, and data requirements in your instructions.

4. Use Reflection:

   • Always ask for a plan before generating code. This helps the AI “think” through the problem and produce better results.

5. Leverage Local Storage for Prototypes:

   • For small apps, use local storage to save data instead of setting up a database. This simplifies the development process.

Key Takeaway for PMs and Founders

Success with AI tools comes down to preparation and precision. Be specific, plan ahead, and take small steps to avoid getting stuck.

🌟 How AI Changes the Role of PMs and Founders

AI tools like Bolt and Replit are transforming the way PMs and founders work. Instead of relying solely on engineers and designers, they can now:

• Prototype Features Independently: Build and test ideas without waiting for development resources.

• Communicate Visually: Use prototypes to share ideas with stakeholders and customers.

• Learn Technical Concepts: Gain a better understanding of how systems work, even without coding expertise.

Colin believes this shift will lead to more collaborative and efficient workflows, where PMs, designers, and engineers can focus on what they do best.

Key Takeaway for PMs and Founders

AI tools empower PMs and founders to take a more hands-on role in product development. Use this opportunity to prototype, test, and iterate faster than ever before.

🔓 Final Thoughts: Unlocking the Power of AI in App Development

If you’re ready to start building apps with AI, try out tools like Bolt or Replit and experiment with Colin’s reflection technique.

For a deeper dive on From Idea to App in Minutes: The Complete Guide to Prototyping with Bolt, check out our Productivity Tapas Post below 👇️ 

or Colin’s course on AI Prototyping for Product Managers, where he covers prompting techniques, technical foundations, and more..

By combining the right tools with the right mindset, you can unlock the full potential of AI and take your app development skills to the next level.

 🎥Watch the full episode here

As promised - I found this great looking addition to Bolt that allows you to edit certain parts of the app much more easily. It’s in beta but you can get access to the waitlist on the link.

📅Timestamps:

• (00:00) Do this before asking AI to code

• (01:13) How to pick the right AI prototyping tool

• (04:40) Useful personal apps that Colin built with AI

• (08:26) Building a habit-tracking app step by step

• (11:39) Always start with the data model

• (19:20) Why explanation before coding works so well

• (25:30) Should you still learn to be technical with AI?

• (29:48) Avoiding the 70% problem and getting stuck on bugs

• (31:38) How to learn to code through AI prototyping

That’s a wrap.

As always, the journey doesn't end here!

Please share and let us know what you would like to see more or less of so we can continue to improve your Product Tapas. 🚀👋

Alastair 🍽️.